摘要
黄酮类化合物广泛存在于植物的各个部位,根据其结构,主要分为黄酮、黄酮醇、双氢黄酮、异黄酮等。对于大多数植物化学工作者来说,解析一个未知化合物都必须经历先确定骨架类型,后确定基团位置这一过程。而前者则需要对该类型的数据规律有充分的认识,否则,就可能导致错误结果。
A set of 64 13C NMR spectra of flavone compounds were examined for information concerning chemical classes by pattern recognition. The Shannon information content for each chemical shift channel was calculated, the 14 channels with the highest information content were retained as a compressed basis set for SIMCA, LDA and KNN pattern recognition. The inherent class structure of the data showed two classes by NLM. Classification accuracy was 90. 91%, 96. 97% and 90. 91% for two classes for SIMCA, LDA and KNN, respectively.
出处
《高等学校化学学报》
SCIE
EI
CAS
CSCD
北大核心
1993年第6期775-777,共3页
Chemical Journal of Chinese Universities
关键词
模式识别
NMR
黄酮类化合物
Pattern recognition, Shannon information theory, 13C NMR spectra, Flavone, Flavonol
